Oncogene (2006) 25, 806–812 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ONCOGENOMICS The elusive multiple self-healing squamous epithelioma (MSSE) : further mapping, analysis of candidates, and loss of heterozygosity

S Bose1,5, LJ Morgan1, DR Booth2, DR Goudie3, MA Ferguson-Smith4 and FM Richards1

1Section of Medical and Molecular Genetics, Division of Reproductive and Child Health, University of Birmingham, Edgbaston, Birmingham, UK; 2Department of Pathology, University of Cambridge, Cambridge, UK; 3Department of Pathology, Ninewells Hospital and Medical School, Dundee, UK and 4Department of Clinical Veterinary Medicine, University of Cambridge, Cambridge, UK

The MSSE gene predisposes to multiple invasive but self- 9q and a shared haplotype was revealed, suggesting a healing skin tumours (multiple self-healing epitheliomata). founder mutation (Goudie et al., 1993). Subsequently, we MSSE was previouslymapped to 9q22–q31 localized the MSSE region to 9q22.3, between D9S197 and a shared haplotype in affected families suggested a and D9S287/D9S1809 (Richards et al., 1997). Our founder mutation. We have refined the MSSE critical continued efforts to refine the MSSE critical region using region (o1 cM, o1 Mb) between the zinc-finger gene new polymorphic markers have been hampered by ZNF169 and the Fanconi anaemia gene FANCC.By ambiguitiesinthemapinthisregion,estimatedtobe genetic mapping we have excluded ZNF169 and FANCC 2.06 cM or 2.2 Mb, 93.34–95.54 Mb from NCBI Build as well as PTCH (PATCHED) and TGFBR1 (transform- 35.1, http://www.ncbi.nlm.nih.gov and Rutgers combined ing growth factor beta receptor type-1) . The linkage-physical map, build 35 (Kong et al., 2004). CDC14B cell cycle phosphatase gene also lies in the Chromosome 9q22.3 is relatively gene-rich and region but screening of the complete coding region harbours several possible candidate genes, including revealed no mutation in MSSE patients. Somatic cell PATCHED (PTCH), ZNF169 (encoding Kruppel–like hybrids created by haploid conversion of an MSSE zinc-finger ), Fanconi Anaemia Complementa- patient’s cells enabled screening of the MSSE chromo- tion group-C (FANCC) and CDC14B. The Gorlin some 9 and showed no CDC14B deletion or mutation that Syndrome gene, PTCH, is a tumour suppressor gene abrogates CDC14B mRNA expression. Thus, CDC14B is (TSG) that is mutated or deleted in the majority of basal unlikelyto be the MSSE gene. We also report the first cell carcinomas of the skin (Gailani et al., 1996; molecular analysis of MSSE tumours showing loss of Holmberg et al., 1996). The ZNF169 gene was mapped heterozygosity of the MSSE region, with loss of the to the interval D9S196–D9S280 (Levanat et al., 1997) normal allele, providing the first evidence that MSSE is a and was therefore considered a candidate. Previously we tumour suppressor gene. have shown that MSSE families share an allele at an Oncogene (2006) 25, 806–812. doi:10.1038/sj.onc.1209092; EcoR1 RFLP in FANCC (Richards et al., 1997) thereby published online 19 September 2005 leaving it a possible candidate. The cell cycle regulatory phosphatase CDC14B regulates p53 (Li et al., 1997, Keywords: skin cancer; tumour suppressor gene; chro- 2000). The map position of CDC14B is ambiguous: mosome 9q22.3; CDC14B; MSSE according to radiation hybrid mapping it lies between D9S196 and D9S287 (GeneMap’99) within the MSSE critical region, but according to the genome sequence (NCBI Build 35.1) it is approximately 0.75 Mb distal to Introduction D9S287, outside the MSSE region. This discrepancy will be discussed later. Individuals with multiple self-healing squamous epithe- Here, we first narrowed down the MSSE interval by ONCOGENOMICS lioma (MSSE) develop multiple invasive skin tumours typing additional markers in MSSE families. Secondly, that undergo spontaneous regression leaving pitted scars, we have analysed possible candidates mapped within or with an age of onset between 8 and 62 years. MSSE shows close to the MSSE critical region, successfully excluding autosomal dominant inheritance, with most affected some of them. Thirdly, we have performed the first families originating in western Scotland. By linkage molecular analysis of MSSE tumours and have detected analysis, the MSSE gene was mapped to chromosome loss of heterozygosity (LOH), providing first evidence that MSSE is likely to be a TSG. Correspondence: Dr FM Richards; Current address: DanioLabs Ltd., 7330, Cambridge Research Park, Waterbeach, Cambs CB5 9TN, UK. Results E-mail: [email protected] 5Current address: Cancer Research UK Institute for Cancer Studies, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK. Refining the genetic map of the MSSE interval Received 11 November 2004; revised 19 July 2005; accepted 25 July 2005; To reduce the MSSE interval, we performed haplotype published online 19 September 2005 analysis with nine new markers. However, due to gaps The MSSE skin tumour suppressor gene S Bose et al 807 and ambiguities in the physical map of this region, we first had to genetically map these markers (D9S119, FBP1, AFM070xb11, AFMa086yf1, AFM203wh8, AFMa350xg1, AFM023xh8, D9S1851, TGFBR1)by typing them in 15 CEPH and three Gorlin Syndrome families with known 9q22 recombinations, along with seven previously mapped markers (ZNF169, D9S280, D9S1816, D9S196, D9S197, D9S287, D9S1809). Posi- tional information from three Gorlin families and six CEPH recombinants (in families 1420, 1421, 13291 and 13292) allowed us to assemble an unambiguous genetic map of the region between D9S196 and D9S180 in relation to the background genetic map based on Genethon, Marshfield and deCODE genetic maps (Figure 1). This mapping enabled us to place TGFBR1 distal to D9S1809, outside the MSSE critical region. Similarly, a key recombinant demonstrated that AF- Ma086yf1 is proximal to FANCC. CDC14B was not mapped genetically because no polymorphisms were known. Our genetic map of these markers is in agreement with the latest assembled sequence (NCBI Build 35.1), with the exception of AFMa086yf1. Our recombination data clearly places AFMa086yf1 prox- imal to FANCC, whereas the assembled genome sequence shows AFMa086yf1 distal to FANCC.We have not been able to find any other reports of genetic mapping of AFMa086yf1. It is important to note that D9S280 and D9S1851 were not given a physical location in the deCODE genetic map (Kong et al., 2002) because the authors were unable to resolve discrepancies between their genetic mapping and the physical draft sequence in this region. The more recent Rutgers combined linkage- Figure 1 Genetic map of 9q22–q31. Marker positions determined physical map (Kong et al., 2004) also does not include from genetic mapping data only (not physical data). The backbone D9S280, suggesting that genetic and physical maps on the left shows the combined data on markers from the Genethon and Marshfield sex-averaged linkage maps and deCODE could not be reconciled. We wonder if there is a genetic map, including chromosomal orientation and the approx- duplication in this region that has resulted in these imate genetic distance of 2 cM. To the right of the backbone is the discrepancies. This draws support from evidence of a genetic mapping data from the present study obtained from key duplicated D9S280 locus >2 Mb distal to the first recombinants in CEPH and Gorlin syndrome families. Vertical lines next to a marker name indicate the position of the marker D9S280, also distal to D9S287 in the annotated genome (e.g. AFM070xb11 is between D9S196 and FANCC). Arrows indi- sequence. cate that the possible position extends beyond limits of the map shown (e.g. D9S119 is proximal to D9S197). Sloping lines indicate the position of markers with respect to each other (e.g. AFM203wh8 Genotyping of intragenic polymorphisms excludes three is distal to AFMa086yf1 and proximal to AFMa350xg1 and AFM023xh8). CDC14B is not included because it has not been genes mapped genetically. The FANCC, ZNF169 and PTCH genes lie within the MSSE interval (D9S196–D9S287) and were thus con- polymorphism (C/G at nucleotide1504–51) showed that sidered as candidates. We have now typed intragenic the C allele segregates with the disease in affected LE polymorphisms in 11 MSSE families known to share a family members, in contrast to a G allele in other haplotype on chromosome 9q. families, thereby excluding PTCH as a candidate. Affected members of LE and BL families do not share the common allele at FANCC intron 2 (CAintr2) and the LE family also does not share the common allele at Haplotype analysis refines the MSSE interval FANCC intron 1 (CAintr1a) (Table 1). Also, affected To narrow down the MSSE critical region, we ana- members of families BL and LE did not share an allele lysed eight additional markers (D9S119, FBP1, with the rest of the families at a ZNF169 intronic AFM070xb11, AFMa086yf1, AFM203wh8, AF- microsatellite (Table 1). Thus, FANCC and ZNF169 Ma350xg1, AFM023xh8, and D9S1851) in the MSSE were both excluded as MSSE candidates. families (Table 1). Affected individuals from all of the 11 Earlier we had failed to exclude PTCH (Richards families previously known to share the 9q haplotype et al., 1997); however, typing of an additional PTCH were found to have at least one allele in common at each

Oncogene The MSSE skin tumour suppressor gene S Bose et al 808 Table 1 Chromosome 9q haplotypes in MSSE families

Polymorphic alleles on 9q13–9q31 segregating with the MSSE phenotype in 11 affected families. Shaded region indicates shared haplotypes. When two possible alleles are shown for an individual, it either indicates heterozygous state or that the phase could not be determined. When more than two alleles have been shown (e.g. in BL and LE), it indicates that all affected individuals within the family did not share a common allele. The relative positions of markers shown in bold have been confirmed by genetic linkage mapping.

of these marker loci, with the exception of the most CDC14B is not mutated in MSSE patients distant families, BL and LE. Affected BL members did Analysis of the annotated genome sequence showed the not share a common allele at D9S119 or AFM023xh8, CDC14B gene to be only 50 kb from marker and affected members of LE did not share the common AFMa086yf1. As discussed above we believe that there allele at D9S1851. is a discrepancy in the genome assembly and that Analysis of the haplotype data from the microsatel- AFMa086yf1 may be proximal to FANCC and thus lites and intragenic polymorphisms suggests several within the MSSE critical region, making CDC14B a possible locations for MSSE: either in the interval candidate for MSSE. We screened all 15 CDC14B exons ZNF169–FANCC, or distal to FANCC close to either in 13 MSSE patients representing five families and eight AFM203wh8 or D9S1816 (Table 1). However, there is a isolated cases: no mutation was identified. We identified recombinant individual in family GE, who is unaffected an intronic polymorphism, a GTG insertion in a six at more than 67 years of age, shows high-risk haplotype GTG repeat (intron 14 ntd À55) in GE4.33 (which did at FANCC, PTCH, D9S287 and distal markers, but not segregate with MSSE in the family), one isolated low-risk haplotype at D9S280, AFM070xb11, ZNF169 MSSE case and normal individuals. and more proximal markers (Figure 2). Assuming that Further, to rule out large deletions that may have this individual, GE6.16, is truly unaffected and is not a been missed in the heterozygous patients during muta- mutation carrier then MSSE is proximal to FANCC. tion analysis, we created somatic cell hybrid clones, each Thus, the haplotype and key recombinant data suggest containing a single from MSSE patient that the most likely position for MSSE lies within the LE2.7, that would allow screening of haploid DNA. All 1 cM distance (o1 Mb), between ZNF169 and FANCC. CDC14B exons were detected by PCR from these hybrid

Oncogene The MSSE skin tumour suppressor gene S Bose et al 809

Figure 2 Haplotypes on 9q22–q31 in part of the pedigree of GE family, showing the unaffected individual GE6.16 who has a key recombination. The black bar represents the high-risk (mutant) chromosome, the white and grey shaded bars represent the low-risk . The numbers beside each bar represent different alleles of the markers indicated at the left-hand side. The markers are in the proximal to distal order from top to bottom. Affected individuals are represented by black shapes and unaffected individuals by white shapes. clones, showing that there was no deletion in the entire patient BH but this tumour retains heterozygosity at coding region of CDC14B. AFM070xb11 and thus does not appear to show an To rule out a promoter mutation, we detected overlapping region of LOH with the tumours from CDC14B mRNA expression in all haploid hybrids SE3.25. However, the markers distal to AFM070xb11 equivalent to levels in diploid lymphoblastoid cells from that were tested were not informative in BH so a second a normal individual and LE2.7. Both isoforms were region of LOH in the MSSE region in this tumour detected (CDC14B2 containing exons 1–14 and cannot be ruled out. CDC14B1, missing exon 13), regardless of the chromo- In all informative tumours from the familial cases some 9 allele present. We concluded that there is no SE3.25 and AR2.4 the allele lost was always the normal mutation, deletion, or loss of expression of CDC14B in allele rather than the allele that segregates with MSSE. MSSE patients. Thus, it is likely that MSSE is a TSG.

Evidence of LOH around the MSSE interval To investigate presence of LOH, six markers were typed Discussion in 12 tumours from five MSSE individuals: four tumours from SE3.25, three from AR2.4, one each We have localized the MSSE locus to the o1 cM region from LE2.2 and LE2.7 and three from isolated case BH. between ZNF169 and FANCC based on haplotype and In seven tumours no LOH was detected; however, LOH linkage analysis in 11 families with a common ancestor, was observed in the remaining five tumours. In all cases including a key recombinant individual who is un- the loss was partial, probably due to contamination with affected at 67 years of age. The corresponding physical normal cells (Figure 3a). LOH across the whole region distance is likely to be o800 kb, however, some was detected in tumour T3 from SE3.25. In three of the discrepancies between the physical assembly of the other tumours loss was detected at some markers but genome and the genetic map of this region still remain. not others (Figure 3b). There appears to be a common A recent assessment of the gaps in the human genome region of deletion distal to AFM070xb11 and D9S280 in sequence showed that the region between 92 and 96 Mb the tumours from SE3.25. This region of LOH is within on chromosome 9 (approximately D9S196–D9S180) the critical region for MSSE identified from the family contains duplicated sequence (Eichler et al., 2004), studies. Analysis of additional markers would be which we believe may have led to some errors in required to identify the minimal region of loss. This is sequence assembly. The order of markers in this region hampered by a paucity of tumour material because confirmed by both genetic and physical maps is MSSE tumours are now treated with cryotherapy rather D9S196–ZNF169–D9S280–FANCC–D9S1816–D9S287, than being surgically excised. There is a region of LOH but there is an additional ‘pseudo’ D9S280 duplicated around ZNF169 in the tumour from the sporadic MSSE locus in the annotated genome sequence more than

Oncogene The MSSE skin tumour suppressor gene S Bose et al 810 is a polymorphism of genome organization (such as a duplication or inversion) that differs between the DNA used for genome sequencing and that used for genetic mapping studies. There is a precedent for this: a polymorphic inversion has recently been implicated in ethnic differences in mapping of a region of chromo- some 8p (Jorgenson et al., 2005). While we have not formally investigated whether MSSE might be caused by a duplication mutation, none of our electrophoretic gels have shown any indication of higher gene or marker dosage in affected individuals than normal individuals, and this type of mutation is unlikely for a TSG, which we believe MSSE to be. We have excluded four plausible MSSE candidates by genetic mapping (FANCC, ZNF169, PTCH and TGFBR1). For another possible candidate, CDC14B, we found no deletions or mutations in the entire coding region, and no mutation that abrogates mRNA expres- sion. The data that excluded deletion and loss of expression was obtained using somatic cell hybrids created from an affected individual who shares the founder genotype at the MSSE interval. This is a powerful method for excluding germline mutations that would otherwise be missed by PCR analysis of hetero- zygous DNA. Thus, while we have been unable to totally exclude CDC14B, it is unlikely to be the MSSE gene. Refinement of the MSSE critical region, together with human genome sequence data has also enabled us to exclude several other possible candidate genes on 9q22– q31. Thus, the genes CORO2A (Zaphiropoulos and Toftgard, 1996), CKS2 (Demetrick et al., 1996), DAPK1 (Feinstein et al., 1995), DEC1 (Nishiwaki et al., 2000), ECM2 (Nishiu et al., 1998), FKHL15 (Chadwick et al., 1997), GADD45G (Suzuki et al., 1999), GAS1 (Wicking et al., 1995), NINJ1 (Chadwick et al., 1998), RGS3 (Chatterjee et al., 1997), TRK-B/NTRK2 (Slaugenhaupt et al., 1995) and GKLF implicated as TSG in colorectal cancer (Zhao et al., 2004), all lie outside the MSSE critical region. There are at least six genes between Figure 3 (a) Example of LOH at D9S1809 in an MSSE tumour. ZNF169 and FANCC, which are positional candidates Genescan gel of samples from SE3.25 showing alleles 5 (130 bp) and 3 (134 bp) in normal DNA (N) and loss of allele 3 in tumour for MSSE: AL133071, FLJ14753 and AF130099 (with T4. Molecular weight marker shows 139 and 150 bp peaks. unknown function), FBP1, FBP 2 (fructose bisphos- (b) Patterns of 9q LOH in 5 MSSE tumours. White circles: tumour phatase genes) and C9orf3 (predicted to be a membrane is heterozygous status. Black circles: LOH in tumour. No circle alanine aminopeptidase). MicroRNA genes have re- either indicates the marker was not informative or no result was obtained due to a paucity of DNA. The marker order confirmed cently been associated with cancer (Croce and Calin, by both genetic and physical maps is D9S197–ZNF169–D9S280– 2005) and a cluster of three microRNA genes lies just D9S1809. The positions given for AFM070xb11 and AFMa350xg1 proximal to FANCC on 9q22 (miRBase http://microrna. with respect to these markers are based on physical maps but have sanger.ac.uk, Griffiths-Jones, 2004). These are worthy of not been confirmed by genetic mapping. further investigation. Identification of LOH around the MSSE locus in tumours, with consistent loss of the normal allele, 2 Mb distal to the first and also distal to D9S287 and we suggests that loss of function of MSSE is likely to be wonder if this duplication may have contributed to important for tumourigenesis. This provides the first errors in physical mapping. AFMa086yf1 is placed distal direct evidence that MSSE fits Knudson’s two-hit to D9S287 on the annotated genome sequence but we hypothesis and is a TSG. It is possible that this gene is have shown by genetic linkage with key recombinants also involved in the development of other tumours. Up that AFMa086yf1 is proximal to FANCC.Itis to 65% of cutaneous squamous cell carcinoma (SCC) intriguing that several groups have found discrepancies have LOH at a region of 9q22.3 but no mutations in in this region but we are unable to provide an PTCH or XPA were found, suggesting the existence of explanation, apart from speculating that one possibility another TSG in this region important for sporadic SCC

Oncogene The MSSE skin tumour suppressor gene S Bose et al 811 development (Ahmadian et al., 1998). Also TSGs for for mutations in CDC14B (Genbank Accession numbers other cancers are suspected to reside between 9q22–q31 AF023158, AF064104 and AF064105). Exons demonstrating (Byrom et al., 2004). It is possible that the MSSE TSG band shifts were sequenced. All exons were screened by SSCP may play a role in these other sporadic tumours. in patients GE4.33, SE3.25, LR4.17, JA (from BL family), MO (from LN family) and eight isolated cases. Then for final exclusion all 15 exons were directly sequenced from GE4.33 and two isolated cases. Materials and methods For screening of CDC14B deletions and expression, somatic cell hybrid clones haploid for chromosome 9 were created from Patients and samples lymphoblastoid cells of LE2.7 and murine E2 embryonic DNA from 161 individuals, including 49 affected and 104 fibroblast recipient cells using GMP Conversion Technology unaffected or at risk members from 11 MSSE families and (GMP Genetics, Inc.). We obtained DNA from four hybrid eight isolated MSSE cases was studied. The nomenclature used lines each containing one of the chromosome 9 (clones 7 and for the families is consistent with previous publications on 10 with one, clones 8 and 18 with the other) that is, two clones MSSE. with normal chromosome 9 and two with the MSSE We also analysed DNA from 15 CEPH families (12, 66, 104, chromosome 9. All four clones were analysed for CDC14B 884, 1332, 1345, 1346, 1377 (two recombinants), 1416, 1420, deletions by PCR of each exon. 1421 (two recombinants), 1423 (two recombinants), 13291, RT–PCR was performed on hybrid clones 7 and 8, 13292 (two recombinants), 13293, (http://www.cephb. lymphoblastoid cells from LE2.7 and a normal individual, fr/cephdb/)) and three families with Gorlin Syndrome (Farndon mouse E2 cells. Primer sequences and PCR conditions are et al., 1994) with known genetic recombinations in the 9q22.3 available on request. region, kindly provided by Professor Sue Povey and Professor Peter Farndon, respectively. LOH analysis of MSSE tumours Normal and tumour tissue regions marked on haematoxylin– Analysis of 9q polymorphic markers for linkage and haplotype eosin-stained slides were dissected from sequential formalin analysis fixed, paraffin-embedded sections. DNA obtained following Microsatellite markers D9S119, D9S197, D9S196, ZNF169, proteinase K digestion (551C/overnight) was amplified and FBP1, D9S280, D9S1816, D9S287, AFM070xb11, AF- compared with DNA from peripheral blood. Ma086yf1, D9S1851, AFMa350xg1, D9S1809, AFM023xh8 Forward primers for all six markers (D9S197, ZNF169, were amplified by PCR using published primers (http:// AFM070xb11, D9S280, D9S1809, AFMa350xg1) used for www.gdb.org) and standard conditions. For AFM203wh8, LOH analysis were labelled with either HEX or 6-FAM new primer AFM203F2 (50-CTG TTA CTC TGG TGA CTA (Invitrogen Life Technologies Ltd, UK). After 25 cycles of CCC A-30) and published primer AFM203wh8.m were used. amplification, PCR products were analysed by 6% PAGE Two CA repeats in intron 1a and intron 2 of the FANCC were using an ABI377 (Applied Biosystems) with Genescan-500 amplified (Savoia et al., 1996). All samples were analysed by TAMRA as internal size standard and Genescan software denaturing PAGE with silver staining. (version 2.0.2) (Applied Biosystems). LOH was scored by calculation of the ratio of the tumour DNA peaks (T1/T2) Analysis of intragenic polymorphisms compared to that in normal DNA to give relative ratio XLOH A 255 bp fragment containing PTCH exon 11 and C/G [(T1/T2)/(N1/N2)]. XLOH ¼ 0 signifies complete allele loss, 1 polymorphism in intron 10, nt 1504–51 (Richards et al., signifies no LOH. We took XLOHo0.5 to represent LOH, 1997) was amplified using primers PTCF5 (50-GTG TTA GGT although other groups have used values as high as o0.8 (that GCT GGT GGC A-30) and PTCR5 (50-CTT AGG AAC AGA represents 30% LOH-containing cells contaminated with 70% GGA AGC TG-30). After NlaIII digestion the alleles were normal cells) (Martignetti et al., 2000). detected as C allele: 63 þ 192 bp; G allele: 18 þ 45 þ 192 bp. TGFBR1 intron 7 polymorphism (G/A at 1255 þ 24) (Gong Acknowledgements et al., 1999) was analysed in CEPH families informative for We are grateful to all the MSSE families for their cooperation. D9S1809 by PCR-SSCP of a 248 bp fragment. Sequencing was We thank Professor Peter Farndon, Professor Sue Povey, performed to confirm results that were inconclusive. Dr Roy Palmer, Dr Colin Arlett, Dr Arjida Woollons, Dr Anne-Marie Gerdes, Sean Humphray and dermatologists Analysis of CDC14B from across the UK for information and samples used in this PCR-SSCP with intronic primers for all 15 exons (sequences study. This work was funded by the Medical Research Council and PCR conditions available on request) was used to screen and Tenovus, UK.

References

Ahmadian A, Ren ZP, Williams C, Ponten F, Odeberg J, Demetrick DJ, Zhang H, Beach DH. (1996). Cytogenet Cell Ponten J et al. (1998). Oncogene 17: 1837–1843. Genet 73: 250–254. Byrom J, Mudaliar V, Redman CW, Jones P, Strange RC, Eichler EE, Clark RA, She X. (2004). Nat Rev Genet 5: 345–348. Hoban PR. (2004). Int J Oncol 24: 1271–1277. Farndon PA, Morris DJ, Hardy C, McConville CM, Chadwick BP, Heath SK, Williamson J, Obermayr F, Patel L, Weissenbach J, Kilpatrick MW et al. (1994). Genomics 23: Sheer D et al. (1998). Genomics 47: 58–63. 486–489. Chadwick BP, Obermayr F, Frischauf AM. (1997). Genomics Feinstein E, Druck T, Kastury K, Berissi H, Goodart SA, 41: 390–396. Overhauser J et al. (1995). Genomics 29: 305–307. Chatterjee TK, Eapen A, Kanis AB, Fisher RA. (1997). Gailani MR, Stahle-Backdahl M, Leffell DJ, Glynn M, Genomics 45: 429–433. Zaphiropoulos PG, Pressman C et al. (1996). Nat Genet Croce CM, Calin GA. (2005). Cell 122: 6–7. 14: 78–81.

Oncogene The MSSE skin tumour suppressor gene S Bose et al 812 Gong YQ, Chitayat D, Kerr B, Chen TP, Babul-Hirji R, Pal A Li LW, Ljungmann M, Dixon JE. (2000). J Biol Chem 275: et al. (1999). Am J Hum Genet 64: 570–577. 2410–2414. Goudie DR, Yuille MAR, Leversha MA, Furlong RA, Carter Martignetti JA, Gelb BD, Pierce H, Picci P, Desnick RJ. NP, Lush MJ et al. (1993). Nat Genet 3: 165–169. (2000). Genes Chromosomes & Cancer 27: 191–195. Griffiths-Jones S. (2004). Nucleic Acids Res 32(Database Nishiu J, Tanaka T, Nakamura Y. (1998). Genomics 52: 378–381. Issue): D109–D111. Nishiwaki T, Daigo Y, Kawasoe T, Nakamura Y. (2000). Holmberg E, Rozell BL, Toftgard R. (1996). Br J Cancer 74: Genes Chromosomes & Cancer 27: 169–176. 246–250. Richards FM, Goudie DR, Cooper WN, Jene Q, Barroso I, Jorgenson E, Tang H, Gadde M, Province M, Wicking C et al. (1997). Hum Genet 101: 317–322. Leppert M, Kardia S et al. (2005). Am J Hum Genet 76: Savoia A, Centra M, Ianzano L, deCillis GP, Buchwald M, 276–290. Zelante L. (1996). Mol Cell Probes 10: 213–218. Kong A, Gudbjartsson DF, Sainz J, Jonsdottir GM, Slaugenhaupt SA, Blumenfeld A, Liebert CB, Mull J, Lucente Gudjonsson SA, Richardsson B et al. (2002). Nat Genet DE, Monahan M et al. (1995). Genomics 25: 730–732. 31: 241–247. Suzuki M, Watanabe TK, Fujiwara T, Nakamura Y, Kong X, Murphy K, Raj T, He C, White PS, Matise TC. Takahashi E, Tanigami A. (1999). J Hum Genet 44: 300–303. (2004). Am J Hum Genet 75: 1143–1148. Wicking C, Breen M, Negus K, Berkman J, Evdokiou A, Levanat S, Chidambaram A, Wicking C, Bray-Ward P, Cowled P et al. (1995). Cytogenet Cell Genet 68: 119–121. Pressman C, Toftgard R et al. (1997). Cytogenet Cell Genet Zaphiropoulos PG, Toftgard R. (1996). DNA Cell Biol 15: 76: 208–213. 1049–1056. Li LW, Ernsting BR, Wishart MJ, Lohse DL, Dixon JE. Zhao W, Hisamuddi IM, Nandan MO, Babbin BA, Lamb NE, (1997). J Biol Chem 272: 29403–29406. Yang VW. (2004). Oncogene 23: 395–402.

Oncogene